Applied Electromagnetic Environment Cognition to Improve a Radio Communication Lin

With the continuous adoption of mobile devices, the cellular wireless capacity is projected to grow at rates that would not be supported in current microwave networks, therefore, a high interest in millimeter-wave bands (30-300 GHz) has risen since they provide the possibility of a larger spectrum allocation compared to the one below 3 GHz. Specifically, the 28-38 GHz band is currently considered to have a high potential to provide fifth generation cellular services. In this regard, field measurements show that viable links, even in non-line of sight conditions, can be established within a radius of 200 meters but these need to be found with the help of directional antennas. This scheme is different from the one used in current cellular systems in which the application of directional antennas is normally applied after the network access is completed and not for the access itself. Channel estimation implemented with hybrid beamforming methods have been proposed as solutions but at the expense of more than one radio frequency chain. On the other hand, compressed sensing channel estimation techniques which take advantage of the sparse characteristics of the millimeter-wave channels giving an accurate representation of the power angle profile and a very much reduced beamforming overhead turns to be sensitive to additive noise. This latter effect can produce an impact mostly in non-line of sight locations where the signal power may not be high enough.;In this work, we take advantage of the site-specific propagation characteristics in which the system is deployed. This knowledge can be exploited at the base station side in the form of a database linked to the position of the user equipment. The location information is becoming more accessible in cellular equipment as a built-in feature with an increasing degree of accuracy.;In the first part, it is shown that a prior selection of the most effective base station antenna directions to deliver the highest possible power value to users located in non-line of sight positions can speed up an access link. Specifically for the 28 GHz band, simulation data from a ray tracing tool applied in specific urban environments indicate that the utilization of arbitrary base station angles in a cell search procedure is not of maximum benefit for the users. Some of the angles are largely or partially blocked by nearby buildings restricting the radio frequency illumination to a maximum power of a lower value compared to what can be obtained from others. A power level benefit higher than 2 dB can be obtained in a defined number of consecutive power measurements, however this performance improvement appears to be conditioned by the severity of the existing street canyon propagation. Moreover, given that the same base station or user equipment discrete angles are repeated individually or together at many different geographical positions, their percentage distribution for maximum power delivery can also be exploited for further improvement. A simplified ray tracing procedure is developed to identify the best base station angles together with their non-uniform percentage distribution. The application impact (two scenarios) of this reduced set of angles is determined by contrasting the improvement in the signal to noise ratio when both the base station and user equipment change their main beam orientations in a random fashion a certain number of trials. Additionally, a comparison of the number of antenna beam switches needed for an access link in a particular antenna alignment procedure is performed. These results may be helpful especially for those systems using analog beamforming.;In the second part, we developed an antenna alignment algorithm exploiting specific propagation knowledge in the form of a database linked to the geographical position of the user equipment. Both cases in which the base station knows the exact and inexact location of the user are considered. We take advantage of the observation that the base station angle associated to the first propagation path serving the user repeat exactly or approximately in the database points located within a circle of radius D around the reported position. This result allows to illuminate appropriately the user even for the case of an inexact position information. The determination of the best BS angle facilitates the application of a narrow beamwidth antenna in the initial stage of the antenna alignment process providing the advantage of an increased starting illumination power and consequently of a reduced total number of alignment steps. Through ray tracing simulations it is shown that the proposed algorithm has similar or better performance relative to a modified classical hierarchical procedure as long as the exact geographical position is located inside a circular area of database points. Moreover the radius D of the circular area can be set to a value a little larger than the maximum error of the positioning system without much variation on the performance. On the other hand, there is no need of a very high resolution database. Simulation results applying maximum position errors up to 10 m and database resolutions up to 4 m, respectively, indicate that the distance between the database points can be increased up to 2m without much impact. These outcomes are sustained when the maximum power level received at the user equipment varies.;In the third part, we expand the application of a site-specific propagation characteristics knowledge to increase a link capacity. Given that a database can have information of all propagation paths associated to a specific position of the user, we extend the algorithm developed in the previous part to take advantage of a second path as long as the total capacity link is larger than what can be obtained with just the first path. The inaccuracy in the position reported forces the selection of a reference point in the database to approximate the main angles of the second path. Ray tracing simulations in three well defined areas show that for the decision rules adopted, the additional capacity gain accomplished with the algorithm is fairly close to the ideal case where the exact user position is known. Given that only one additional step is needed for the base station to communicate the angle information of the second path to the user, a database dependent alignment process remains faster relative to a comparable modified classical hierarchical method. Additionally, we examine the effect of the database resolution on the performance of the proposed algorithm for static and variable conditions.